Do the 5.1 Stereo Headphones Really Work?

Tamor asks: "Zalman, the company behind some extremely high quality PC noise-reducing products are now selling real 5.1 surround sound headphones. The surround effect is achieved by placing 3 drivers in each ear-piece. As a geek-with-young-family this product's pushing all the right buttons for me, it looks cool, and means I can finally achieve surround sound without waking the kids. Or does it? I was sure that to place a sound spatially your brain relies on the delay between hearing the sound in one ear and then the other. If your left ear only hears the left 3 channels, and your right ear only hears the right 3 channels isn't this making it more difficult for spatial placement to happen? Do you know if/how these are achieving surround effect if each ear is only hearing half of the audio field?"

Comments

Ask Slashdot? What about Ask the Manufacturer?

[wolf-]

I'm missing something with this category.
Why not call the manufacturer and ask them how they do it? Maybe get a set from them to demo and test. See if YOU can hear the difference.

Physics Problem

[WyerByter]

To my understanding, your ear places sounds spatially by volume. It sounds louder in the closer ear.

Beyond that, unless you have a really big head, the difference in arrival time to each ear is less than a microsecond. That is surely too small for your brain to comprehend.

Re:Physics Problem

[pbox]

Beyond that, unless you have a really big head, the difference in arrival time to each ear is less than a microsecond. That is surely too small for your brain to comprehend.

No it is not. Strange but true. You can always tell the direction (not just left-right but any degree in 3 dimension) where sounds come from (true only for tone above 100Hz or so). Therefore your ears/brain can somehow decipher the minite differences in sounds arriving to your ears.

However, back to the topic. You have 2 ears, therefore 2 speakers are enough to create a complete 3D soundscape. The 5.1 headphones are pure gimmick. You are better off spending some money on a decent pair of 2 speaker headphones, like AKG/Grado and my personal favorite, Sennheiser. If you like music, you are much better off spending $500 on a pair of headphones, than spending the same on 5.1 speakers.

There are some (classical) recordings out there that are done using a fake head, with mic in place of the eardrums. When using in-ear-canal headphones (think Shure / ER) you are placed in the sound environment exactly like where that head was. I belive it is called aural recording, but please post reply if you have correct info.

BTW, comp buffs, EAX by Creative is a model, which creates the 3D sound enviroment. It is a model on how our ears work and you can think of it as a 3D modeling, where you specify the 3D coordinates and the enclosing space (and types of walls), and the system outputs a L/R (or 5.1) signal which tries (with various degree of success) to place it in the right spot.

Re:Physics Problem

[Otter]

There was a famous neuroethology experiment with barn owls. They have asymmetrical ruffs on their ears, one pointing up and one down. Sounds have a different volume in each ear depending on the altitude of the source.

But, it was also showed (by putting headphones on them, playing a mouse sound and watching how their heads moved) that they use volume to determine altitude and time offset to determine bearing. So it's definitely possible -- although I have no idea what system human perception uses for the same problem.

Re:Physics Problem

[Jerf]

However, back to the topic. You have 2 ears, therefore 2 speakers are enough to create a complete 3D soundscape. The 5.1 headphones are pure gimmick.

The 5.1 headphones would be pure gimmick, if we had been able to work out the sound transformations for convincing the brain a sound is coming from a given direction.

AFAIK, there has been progress in the field but it has hit a wall, and all the demos I've ever heard impart a very synthetic characteristic to the sound vs. the original source. (And I'm not speaking as an "audiophile"; the degradation in sound quality is clearly audible to me.)

The headphones can off-load these computations that are so freakishly complex we still can't do them onto reality itself, since "reality" remains better then any algorithm we've put together yet and doesn't sound synthetic.

Now, I've never used these or even heard of these, but I can easily believe that they are more then a gimmick at our current levels of understanding of sound spatialization. Nor would I expect two-speaker setups (headphones or otherwise) to match these any time in the forseeable future.

There are some (classical) recordings out there that are done using a fake head, with mic in place of the eardrums.

This goes to prove my point. Normally "preprocessing" sound before it gets to production is an anathema to a sound engineer; there's virtually nothing you can do to improve the sound while recording it, except record it with higher fidelity. "Aural recording" (if that's the name) is an exception, because you can't add that in post-processing, no matter what inputs you provide yourself. If it was something that could be added in post-processing, the sound engineers would insist on doing so to maintain maximal control over the sound.

Re:Physics Problem

[pbox]

Now, I've never used these or even heard of these, but I can easily believe that they are more then a gimmick at our current levels of understanding of sound spatialization. Nor would I expect two-speaker setups (headphones or otherwise) to match these any time in the forseeable future.

I agree.

So to summarize all this:

1. If the recording is mode with the fake-head, it is best to use 2.0 headphones / in-ear-canal or otherwise.
2. Rest of stereo audio sources are best with a 2.0 headphones
3. Computer generated sounds (especially FPS) best with 5.1 headphones (no or less calc involved)
4. DVD-Audio, SACD 5.1 sources are best with 5.1 headphones, IF are not remastered from a stereo source, but rather are recorded with 5 mono microphones.

Does anyone can improve / extend on above, please post.

Re:Physics Problem

[Reducer2001]

Nice info in parent post. The recording technique is called binaural recording.

Re:Physics Problem

[mleczko]

Well, actually, you only need two channels for surround sound. There are quite a few factors to define where a sound source is located. As far as the horizontal plane is concerned, the location is determined by two factors. First the time difference between the arrival at the ears is taken into account (makes for a astonishigly small time scale, especially with higher frequencies, but the brain can handle it). Then, the volume difference is evaluated: If a sound source is to your left, the signal is louder in your left ear then in your right. As far as vertical position is concerned, the form of your outer ear is relevant. Dependent on the position of the sound source, different frequency bands are attenuated or amplified. These are the so called HRTFs (Head Related Transfer Functions). Using this information, you can filter your sound sources with the according transfer functions and get a really realistic result. I once was able to try out such a system as part of a course here at university. It simulated 5 sound sources in a room and there was a head-tracker mounted on the headphones. So if you turned your head left, the drums would become louder. Pretty cool stuff! :-)

Well calculated delay

[forsetti]

The Headphones are "smart" enough to create an appropriate delay, per channel, to cause that spatial effect you refer to.

google?

[iamjim]

reviews found at:

bigbrui.com, overclockersclub.com, modthebox.com, pcextreme.net, Tom's Hardware, AnandTech.com, etc...

Read CPU magazine for a review

[phoenix_rizzen]

This month's CPU magazine has a review of these headphones. Don't recall the specifics, but they received a good review. The reviewer found them to be much better than stereo headphones during gaming sessions as you could hear sounds from all directions. But the sound quality for DVD movie playback wasn't so hot.

There might be a copy of the review on their website (no I don't have a URL, use a search engine).

a friend has these

[gyratedotorg]

a friend of mine has these. i havent tried them yet, but he's been raving about them.

Not to be snarky:

[attaboy]

But the Zalman product page that you linked to in your post had links to several online reviews. Were those insufficient? I found them to give me all the information that I would need to make a $40 purchase...

www.rbmods.com

www.hardextreme.org

http://www.fastlanehw.com

www.itpro.no

www.hardware-testdk.com

ohls-place.com

Two Ears

[sabNetwork]

Where did you find those other three ears? Please, I'd love to find out.

It's a gimmick, christ. You only have two ears; it doesn't matter where the sound is coming from. Direction is simulated by the recording, not the headphones.

--

Sony alternative

[HeroicAutobot]

Sony also has the much more expensive MDR-DS5100 and the still even more expensive MDR-DS8000.

I've been very tempted by these, but haven't been able to find many reviews. (I haven't looked for a few months though. Maybe there's more information available now.)

You know what?

[Lord Bitman]

These guys who make the headphones, they sort of do this for a living, so they probably know more about it than you. That is: Anything you can come up with in the first five minutes after hearing about the idea, they rely on already having come up with.
This isnt something that somebody decided one weekend would be neat, and so slapped three headphones together with duct-tape and started talking to magazines. They developed, designed, tested, talked to various manufacturers, looked into methods of distribution. Do you think that in all that time, nobody would have considered how surround sound would be best implimented in a pair of headphones?

Editors need to stop accepting stories with these bullshits tacked on. If you want to make a completely uninformed comment, post a comment after[if] the article is accepted.

Re:You know what?

[Elwood P Dowd]

Ah, yes, the engineers may have thought about it, decided that it would be too expensive to implement, and then done something that doesn't work well but only costs $40 rather than $600 for the competing Sony product.

So, marketting or otherwise, his question is worthwhile. His question can be answered by a myriad of reviews available on the subject, but that's not what you seem to be talking about.

Easy

[wishus]

Just because the right can is on your right ear doesn't mean it can't play something from a left channel. There are three drivers in each can, remember? Even if there weren't, you could still mix the left channel into the right can at the appropriate delay and volume.

Spatialization can work with headphones

[esm]

Sorry, I have no experience with these headphones. But: about 10 years ago, I volunteered as a test subject for some experiments done at NASA Ames (Mountain View, CA, USA). One of them was for a device called the Convolvotron. It's a thingy (sorry for the technical term) which takes multiple sound sources and "localizes" each one so it sounds like it's coming from a different place. It worked incredibly well with only two speakers. The big problem was distinguishing between straight-in-front and straight-behind. With headphones and human ears, I suspect that's just a darn difficult problem. But side-front, side, and side-rear were very easy to differentiate.

Although the tests took place in a sound chamber, they were kind enough to give me a demo tape -- and this tape is amazing. They demo about 5 different voices (simultaneous ATC conversations), both flat and spatialized. Flat, it's impossible to differentiate them. With the convolvotron, it was possible and easy to track each conversation separately. Each one sounded like it came from a different place.

This was early 90s. Processing power has certainly increased since then. It should be possible, and relatively cheap, for someone to use Convolvotron-like technology to convert a 5.1-channel signal to spatialized L-and-R ones for use with regular headphones. There shouldn't be a need for special headphones.

Lots of Google hits for "Convolvotron". Enjoy.

Re:Spatialization can work with headphones

[jhoffoss]

Actually, the straight in front and straight behind thing is related more to the human ear/brain/central sound processing unit/whatever. If a sound takes exactly or almost exactly the same amount of time to travel to both ears, there's no way for the brain to determine a direction without echoes or reverberation or something.

If you close your eyes and have someone snap their fingers directly behind or directly above your head, you probably will not be able to determine quite where it's coming from.

Note: this info is based on one intro course I took my freshman year five years ago, and I'm not a doctor/medical student or anything, so I may be off...

cues

[ballpoint]

Left-right stereo has been here a long time and it works wonders with headphones. No doubt about that.

And since any sound arrives at your two cochleas, it must be possible to simulate any sound position just by exciting your two ears, preferably with in-ear phones.

But I have a hinch that cues about whether a sound is at the back or front come subconsciously from:

1. Turning your head and registering the changes in sound.
2. Echoes and reverb. This only works if you know and 'feel' the room. (*)
3. Changes in frequency response due to the structure of your ears. This only works for sounds you know.

As the headphones are fixed to your head the first, and probably the most important, cue disappears. The room where the sounds were recorded does not match the room you're in, so the second cue disappears. And finally you will be listening to new, unknown sounds. There goes the third cue as well.

But in true /. fashion, I'm posting this without actually having experienced 5.1 headphones with more than one speaker on each side. I'd like to try though.

(*) While I'm listening with isolating in-ear buds, it is strange that the sound changes dramatically the moment I enter a building from the outside. Hard to explain by reverb and echo as there is little sound leakage from the buds to the outside and vice-versa.

Hearing

[DarkDust]

I was sure that to place a sound spatially your brain relies on the delay between hearing the sound in one ear and then the other.

Yes, this information is used for left/right locating. But AFAIK (IANAES, I am not an ear specialist) also interference caused by sonic reflections from your shoulders are needed for locating whether a sound comes from above or below. I don't know how the distinguishes front/rear locating, though.

Sennheiser .. but still no subwoofer

[dk.r*nger]

I tried a pair of Sennheiser headphones some five or six years ago.

I believe they cost about $600 or even more, and they had really great sound. I don't have much experience in headphones, so I'm not sure if this basically would apply to any $200+ set... ?

Anyway, they lacked one big thing: The subwoofer. Half the surround experience is the feeling of the ultra low frequency in your stomach, and earphones just wont do that.

3d sound with 2-speakers

[TrebleJunkie]

Many many moons ago, when I was doing video production work, I received a sample CD from an audio library collection billed as "3D-sound".

I don't know how the stuff was recorded, but it was recorded such that you really could localize the sound, in space, in 3 dimensions, from regular ol' stereo headphones. The most memorable tracks on the CD was of someone getting a haircut. You could hear *where* the scissors "were" around your head. You could tell where the hairdryer was blowing. Not just left-or-right, but *around* your head. The stuff was amazing.

I'm guessing that not just volume and left-or-right determines where you hear things, but phase as well.

But, anyhoo, the point being that you can very likely achieve good surround-sound sounding stuff with just one speaker per ear, and not three.

Re:3d sound with 2-speakers

[mnbjhguyt]

this kind of effect has been around for quite some time (best known example is the beginning of the final cut album from pink floyd) (hint: google for holophonia or holophonics)

iirc it's actually very simple, the sounds were recorded using a dummy head with two mikes where the ears would have been

Something I would like to point out

[ewhenn]

In 5.1 the ".1" is a subwoofer. These headphones can't possibly be 5.1

No.

[3-State Bit]

I was sure that to place a sound spatially your brain relies on the delay between hearing the sound in one ear and then the other.

Knowing nothing about human hearing we can almost rule out this conjecture. Noise travels at about 761.207051 mph and your ears are about a foot apart.

That means there is a difference of 895.706603 microseconds between when the first ear would hear the sound and when the second one would.

This is 1/1116th of a second, meaning that if your brain 'ticks' subconsciously at anything less than 1100 hertz its timing would be too coarse to catch this minute difference.
The brain, in fact, ticks a couple of orders of magnitude slower than this, and moreover the theoretical maximum a single neuron can tick is 2000 hertz, so there would have to be ~0 ms delay in signal propagation between neurons, and the signals would have to make a straight line from each ear toward the area in which the signal is to be processed in order for comparison to occur together with pertinent timing information. (The brain, of course, is not so precisely wired that it could take into account some kind of fixed minute differences in timing among various input sources.)

So we can rule that out. The next idea continues with your implicit assumption that each ear is, logically, a fixed point of input, with the brain reconstructing all spatial information. (Ears, in fact, have a complex set of ridges precisely because they do convey spatial information)

But if we thought of ears as mere fixed points of frequency/amplitude sampling, we might be tempted to think that all spatial information is reconstructed from minute differences in amplitude -- the ear nearer the sound source would hear it more loudly. We can also eliminate this conjecture because the two spheres of possible sound location a given distance from each ear intersect not in one point but a whole arc of possible places. What I mean is, if all your brain knew is : "Ear 1 hears source at A loudness and ear 2 hears source at B loudness, and ear1 is at (x1, y1) and ear2 is at (x2, y2)", then, together with information about how sound loses amplitude with the square of the distance it travels and inversely with the frequency (assume the pertinent natural laws are hard-wired), it could produce the fact: A-ha! The source must be 10 feet from ear1 but 10.23 feet from ear2.

The problem is, there is not ONE point that fits those descriptions, but an infinitely many.

If your ears were just input points, then, if you start playing a sound file on the computer in front if you, it should sound the same with your eyes closed now as it would if you turned around and heard it from behind: Each ear hears an equally loud sound, only now from behind instead of in front. The problem is, you can tell that it's from behind and not from in front of you! (Try a double-blind test if you're not sure -- place one speaker dead in front of you and one speaker an equal distance dead behind you, write a script to randomly play either full left or full right balance, close your eyes and listen to the random tests; you'll always be able to tell where the sound source is coming from.).

Okay, so now we've long-windedly debunked the naive assumptions about how the brain might reconstruct spatial information. How does it?

Beats me.

Re:No.

[Anonymous Cow herd]

Your basic mistake is imagining that human sensory input is clock driven rather than signal driven. Your auditory cortex isn't out there "polling" your ears to see if you're hearing anything, rather, the sensors in the ear signal the cortex once a sound (input event) is detected. Also, while the number of points that match your two 'ear judged' distances are infinite (but for all practical purposes, it's finite, this is a variation on Xeno's paradox) your ears are directional, due to their shape (this is also how you can tell difference between sounds in front and behind)

Re:No.

[Frans Faase]

This argument is completely of the mark. The brains does contain specialized areas for detecting the delay. For low tones the spikes produced by the detecting hair cell, match the wave front. These are than transported to an area in the brain where there is a line of cells where the signals from both ears are at opposite ends. The cells where the signals arrive at the same time (depending on the delay caused by the spike to travel through that cell in the line) produce the strongest response and determine the direction from which the sound originates.

I'd go for 5.2

[cr@ckwhore]

Personally, I'd wait until version 5.2 because we all know that .0 and .1 releases are unstable, and you certainly wouldn't want your ears falling off.

-- disclaimer: This absolutely the most retarded post I've ever made.

I don't believe it's just the ears...

[stvangel]

I've worn hearing aids over the last 20 years or so, and I've come to the realization that sound and perception is a lot more than just what the ears hear. My hearing isn't that bad, but it's difficult to get along without them.

The first thing you have to realize, is it isn't just what your ears hear, it's the vibrations that you feel all over your body that affect your spatial perception. Think about it, if spatial perception was just the difference in sound arrival time between each ear there should be no way for you to tell whether a sound is coming from in front of your or behind you. Yet it's obvious when you close your eyes what direction it's coming from.

All sound is is air pressure changes. Your entire body is a hearing device. Your body can feel the sound waves hitting your front and back and can deduce direction pretty easily from that. Your skull makes a very good resonating cavity to collect and amplify these vibrations. Just go to a rave or live concert and feel the vibrations. One of the tests they always give me is bone conduction, where they put two transducers on my skull behind my ears. I can hear the sounds they produce almost as well as if it was audible coming through my ears.

I doubt the three transducers in each ear is worth the effect. I'd think it'd be much more successful if you had some sort of a band all the way around your head that would vibrate your skull from various directions. You need your ears for Clarity, to understand what you're hearing; but your skin and body can handle sound sensations. Even completely deaf people can sense sounds and directions through the vibrations.

I'd think true 5.1 is just unachievable through a simple pair of headphones, no matter how many in each ear. Granted it may still sound good, but still not as good as a good set of speakers and a subwoofer.

For the smart-asses, the answer

[Tamor]

I found out the actual answer to my question, and no it isn't on Zalman's site or in the reviews, and yes I expect that they did think about this before putting a product out. The answer is that the pinna (the outer part of the ear) catches the sound and funnels it down to the ear-drum. The folds and curves of the pinna alter the waveform of the sound as its funneled, and this happens in different ways depending on the direction in which the sound enters the pinna. The brain picks up those differences and is able to tell whether a sound originated in front, behind, above, below etc. So that's how you're able to spatially place a sound you can only hear in one ear. Neat.

Re:Ask Slashdot? What about Ask the Manufacturer?

[Dun Malg]

I'm wondering where the .1 is in 5.1. It's typically the discreet channel, widely known as the subwoofer channel.

Strap on kidney belt that uses a solenoid to punch you in the gut with every bass thump? (probably not)

I imagine the bass channel is piped in equally to both ears along with the center channel. The reason they still call it "5.1" is probably to indicate that it takes 5.1 audio as input.